This ROI is very large and its shape is not homogeneous, therefore you might be averaging out your effect of interest if it localized more precisely. You need to explore your data interactively to understand if you should be using a smaller scout.
Can the scores extracted from scouts time series (shown in the screen capture) be used in any other statistical software to correlate with clinical variables?
Yes. This "matrix" file is readily accessible as .mat file in the database, or can be exported to a text file. Look in the "File" section of the popup menu for this file.
What the unit of measurement?
The same as the ones used in the source file, therefore it depends on the method that you used:
https://neuroimage.usc.edu/brainstorm/Tutorials/SourceEstimation#Measure_.5BTODO.5D